Operations Research

Activity of the research laboratory

Optimization is the most important basis of modern decision making. The main objective of the Optimization Research Laboratory is a constructive participation in the worldwide development of this, relatively new, branch of science. Optimization is based on mathematics and computer science. It also pays attention to the utilization of new developments in hardware architecture. As a result, the activity of the Research Laboratory extends to the following areas:

  • Theoretical, algorithmic, complexity and computational methods research covering the core areas of optimization, with special emphasis on large scale linear, network, mixed integer, conic linear, smooth convex, nonlinear and stochastic optimization.
  • Theory and practice of optimization modelling.
  • Modelling uncertainty, sensitivity analysis.
  • Design and implementation of robust optimization algorithms.
  • Approximation algorithms, including well founded heuristics.
  • Optimization on distributed architecture, including computational grids.
  • High accuracy optimization (special but important application areas).
  • Enhancement of implementation technology of optimization software (how to implement efficiently theoretically correct algorithms).
  • Application projects, including VLSI design, nuclear core reloading optimization, radiation therapy treatment optimization, electricity portfolio optimization, engineering design optimization.

Research results

The leaders of the lab are also world-wide leaders in the theoretical and practical development and enhancement of the two main families of solution algorithms: the simplex method and interior point methods. Both are aimed at the efficient solution of large to very large scale optimization problems. Their achievements have been incorporated in the most successful commercial software packages. Further results include network optimization, mixed integer, quadratic, conic quadratic, semidefinite and smooth nonlinear optimization. Duality theory and computational methods are developed as well.
Among the important results in the simplex method it is worth quoting the creation of a completely new dual simplex algorithm (phase 1 and phase 2) that is highly efficient in itself but even more so if used as the computational engine in the solution of mixed integer problems. Further significant improvements have substantially increased the algorithmic and numerical capabilities of both the primal and dual simplex method. They have enabled the solution of very large and complex problems that were unsolvable so far.
The modern age of interior point methods have started in 1984 with Karmarkar’s epoch making paper. Since then hundreds of variants of polynomial time IPMs have been developed for linear, smooth convex, conic quadratic and semidefinite optimization problems. The target following concept, the self-dual embedding model and novel ways to do sensitivity analysis are among the notable contributions. Interior point methods not only revolutionarized the theory of optimization, but allowed us to develop powerful IPM based optimization software. Nowadays IPMs are part of all leading optimization packages, and are dominant for conic linear optimization.

Introduction of the heads

thumb maros-istvanIstván Maros, mathematician (Eötvös University, Budapest 1964) is a professor of operations research (OR). He received doctorate (1982) from Eötvös University and CSc (1981) and DSc (2006) from the Hungarian Academy of Sciences. Before joining the University of Pannonia in 2006, he was a professor at the Department of Computing, Imperial College London for 11 years. His research interests include linear and nonlinear programming, network optimization, computational techniques of optimization, implementation technology of optimization software, parallel optimization and applications of OR. He has been the chief architect of 12 optimization systems over 30 years each of which represented the state of the art of its time and built scientific achievements of the author. He is the founding co-editor of Computational Management Science (a Springer journal) and associate editor of several international and domestic scientific journals and book series. He is also a member of several international professional organizations.

thumb terlaky-tamasTamás Terlaky, mathematics (Eötvös University, Budapest, 1979) is a professor of operations research (OR). He received doctorate (1981) from Eötvös University, Budapest) , CSc (1985) and DSc (2005) from the Hungarian Academy of Sciences. He has previously taught at Delft University of Technology, Netherlands, McMaster University in Canada, Lehigh University, PA, USA. At McMaster he also served as the founding Director of the School of Computational Engineering and Science. Founding editor-in-chief of the journal, Optimization and Engineering. Terlaky has served as associate editor of seven journals, conference chair, conference organizer, and distinguished invited speaker at conferences all over the World. Member and former chair of numerous professional organizations, and Fellow of the Fields.

the project is supported
hungarys renewal
szechenyi plan
Copyright © 2017 - University of Pannonia, Faculty of Information Technology

This site uses cookies. By continuing to browse the site, you are agreeing to our use of cookies. Find out more from cookie.